EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Multi-Layer Security-by-Design Approach to Fraudulent Account Detection on Social Media: Integrating Machine Learning and Platform Governance

Ankit Agarwal

American Journal of Technology · 31.05.2026 · 5 (7) · 15-37

DOI: 10.58425/ajt.v5i7.549 ↗

Kurz gefasst

Fachartikel in der Zeitschrift American Journal of Technology, online erschienen am 31.05.2026. Verfasst von Ankit Agarwal. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Aim

This study aims to analyze the challenges of detecting fraudulent social media accounts as a security-by-design problem and to examine the integration of machine learning, platform governance, enforcement, and compliance in addressing the proliferation of fake accounts, including sock puppets, bots, cyborg accounts, compromised accounts, CPNs, and AI-generated fake identities.

Methods

The study proposes a multi-layer detection approach that integrates metadata, behavioral, network graph, and content-based signals into a comprehensive fraud risk score. The framework incorporates gradient-boosted models, graph-based detection, natural language processing, image analysis, graduated enforcement, audit logging, safe rollout mechanisms, and compliance-based controls. The empirical design comprises 250,000 accounts, including 190,000 legitimate accounts and 60,000 fraudulent accounts, supported by 12 million activity records and 3.5 million network edges.

Results

As the study introduces a framework and does not conduct full-scale experimentation on social media platforms, the reported values are projected benchmarks. The proposed multi-signal fusion approach is anticipated to achieve an AUC greater than 0.92, a detection rate of 90–95%, and a false-positive rate of 2–3%.

Conclusion

The study demonstrates that integrating machine learning, governance, enforcement, compliance mechanisms, and multi-signal analysis into a unified framework provides a comprehensive approach to fraudulent account detection on social media platforms. Recommendation: The paper recommends that trusted platforms incorporate detection, governance, privacy, and accountability measures into their routine account management and content moderation processes.

Zitieren

APA

Agarwal, A. (2026). A Multi-Layer Security-by-Design Approach to Fraudulent Account Detection on Social Media: Integrating Machine Learning and Platform Governance. American Journal of Technology, 5(7), 15-37. https://doi.org/10.58425/ajt.v5i7.549

Vancouver

Agarwal A. A Multi-Layer Security-by-Design Approach to Fraudulent Account Detection on Social Media: Integrating Machine Learning and Platform Governance. American Journal of Technology. 2026;5(7):15-37. doi:10.58425/ajt.v5i7.549

BibTeX
@article{Agarwal2026,
  author = {Agarwal, Ankit},
  title = {A Multi-Layer Security-by-Design Approach to Fraudulent Account Detection on Social Media: Integrating Machine Learning and Platform Governance},
  journal = {American Journal of Technology},
  year = {2026},
  volume = {5},
  number = {7},
  pages = {15--37},
  doi = {10.58425/ajt.v5i7.549}
}
Kostenlos starten →